Monday, June 30, 2008

Awareness, Interaction, Engagement – What are these and How to measure?

These terms are used so often in online advertising but what exactly do they mean to an advertiser and how to measure them?

Awareness as the word suggests is when the consumers get aware of the brand. This is the first touch point of the brand with the consumers. This could be done via various channels – Online, TV, out of home (billboards), Search etc. Increase in brand awareness overtime can be measured with pre-post analysis, reaching out to consumers and finding the % awareness in the marketplace and after the campaign doing a similar survey again. The Ad-recall is also a good metric to measure awareness.

Interaction is when a consumer interacts with the brand – this could be mouse-over a rich media banner or calling in for a sweep stake or even clicking on a banner. Some of the common metrics to track this is click-through rate, interaction rate, # of calls received etc.

Engagement should be when the consumer really engages with the brand and goes "Beyond" the basic steps to know more about the brand and tries to build a relationship. Some examples could be signing up for a newsletter, browsing through the site in detail, downloading a widget etc. The deeper the level of engagement the more engaged audience or consumers the brand has.

Thursday, June 26, 2008

Calculating Weekly Goals

It is important to calculate weekly goals to accurately keep track of the campaign performance and be able to make changes to the campaign if it is performing below expectations. One of the most common ways to calculate the weekly goals:

Using Historical Data – If a similar campaign has been live in the earlier, then that previous campaign's results can be used to directionally predict the metrics of the upcoming campaign. Sometimes, even using an x% increase year over year or quarter over quarter can be used. One more factor to be considered in the seasonality of the product/service. If the upcoming campaign falls in a higher or lower seasonality curve, then it should be taken into consideration.

After the actual campaign goes live, the forecasted numbers should be validated with the actual and learnings should be applied during the next campaign to reduce the discrepancies.

Monday, June 23, 2008

All consumers are not equal

How do you define the value of a consumer is who visits a website? If someone visited a single page more valuable or someone who visited multiple pages much more valuable.

I guess it depends on page content. If the single page that was visited was one of the most valuable pages on your website, then that one page visitor should be the most valuable consumer on your website. Such pages could be the core product information or landing page. One more example could be the "Submit" more information page i.e. if the consumer takes some action at your website. If the consumer shares some personal information like email, phone, address then this is a perfect time to start a relationship marketing programs. Ask them what they need and provide them…

A consumer could also visit multiple pages which might not be so valuable and as they would not show stronger engagement with the brand or the product.

In the direct response model, especially with an e-commerce engine in place, a funnel based measurement system works very well to evaluate the value of the consumers. The further down the funnel a consumer is the more valuable he/she is. However, in the case of branding websites, there is no linear path that a consumer is supposed to follow. Thus, in the branding websites each pages needs to be assigned a value and then evaluate the value of your consumers.

Wednesday, June 18, 2008

Click Sequence Tracking – Search, Newsletter, Display, Search

It is very important to keep track of the click sequence before conversion while using more than one channel for a particular campaign. This is very important to apply correct attribution to each media and also to understand how all the different media channels are helping or cannibalizing each other.

The overlap in Search and Display is very high; I have seen numbers as high as 33%. This number could be even higher for products if there is very long research phase like buying a higher priced item example laptops.

There are some advanced reports in DoubleClick's DART (Exposure to Conversion) report which provides the last 10 clicks (by site, placement, time) which helps understand which publisher engaged the most consumers for the first time and then which publisher led to most conversions with least number of other publishers coming in between the first click and conversion.

Some other cookie based systems can provide overlap across all channels – email, search, display, affiliates (DART cannot provide this as DART tracks only Search and Display).


Monday, June 16, 2008

Tracking DRTV using Site Analytics

With the advances of web analytics technology, it is also possible to track the impact of DRTV using the web analytics tools like Google Analytics, WebTrends, Hitbox etc.

Usually, the DRTV ads have a unique URL associated with the TV ad, where the advertiser intends the consumer to go to. So, tracking the number of visits to this page would help understand how many consumers the campaign is driving. Using web analytics tools, funnels can be created to track the complete path of the consumers if the campaign involves making a purchase or registration.

One more interesting thing can be done at the call center is giving the customer service representatives a unique URL to open in their browser whenever they receive a phone call, and then finish the process online – tracking this funnel would help understand the complete drop-offs and help make process optimizations.

Sunday, June 15, 2008

Measuring the effect of Marketing Efforts – Part II

One more way of measuring the effect of marketing efforts is to do a pre marketing and post marketing efforts analysis. Trending one of the key metrics like conversion rate, ROAS, click-through rate etc. If there is an upward trend, it shows that the marketing efforts have paid off however if the metric is trending downward this is not a good indicator.

Upward or downward movement is a very good indicator but other factors should be considered in making the right decision. The trending could be due to other factors like competitive effect, seasonality, etc.

Thursday, June 12, 2008

How to measure lift from any marketing effort?

It is very important to measure the lift from the marketing campaign. This is essential to understand how much more products or consumers are converting due to the marketing/advertising efforts.

A basic test/control methodology should be used to find out the lift in conversion rate due to advertising. This is easy to understand with an example of the grocery store - how likely is someone to buy a product if he/she has not seen any advertising and then how likely is someone to buy a product if he/she has seen advertising. The % difference between the two is the lift from the advertising.

In simple words you found out the purchase intent without the advertising (Control group) and then with the advertising (Test Group). The % difference between the two provides the lift.

How easy is it to implement or execute this?

For Online marketing, it is not hard at all. Comscore has a panel of over 1MM consumers whom they monitor their online behavior. Their solution creates a test/control group based on Ad exposure and conversion.

DoubleClick also has a cookie based solution to segment consumers who saw the ad and those who did not and then calculating the lift due to advertising.

For offline marketing, there are mathematical models which have been used in the past to calculate the lift.

Monday, June 9, 2008

Cost of having the Website Down:

Today I wanted to purchase a product from which I had added to my wishlist yesterday.
After going through the complete process, when I was about to check out, I get this error message that the website is down. I have been using Amazon for years, and I was shocked.

My mind just started thinking, if this site stays down for even 20-25 mins, how many orders will be missed and how much revenue Amazon will lose. Even if we assume they get 100 orders/min with an average order value of $50.00, they would loose $5,000/minute which would translate to about $100,000 - $125,000 in 20-25 mins.

It's not only just revenue but consumers getting frustrated with the situation. Many people just order pretty much everything (clothes to diapers to electronics) from I think pretty much they have the complete market share but however if there was a new consumer on the site during this time, it would be a bad consumer experience and Amazon might lose a consumer. Calculating the consumer life time value would help understand the value of a new consumer.

Moral of the case - Keep a close eye on your e-commerce engine and fix it ASAP.

Bought Display and Search Media, Online sales should go up: Really?

A lot of advertisers tend to think in these terms. People think I bought media, people are looking at my ads, they are clicking on them as well, my sales should go up. Really? Why? In most of the cases it will or should go up, but sometimes it will not. Why? Have you checked your website analytics, i.e. how is the consumer flow? Are there are places where people are dropping off a lot? If yes, that thing needs to be fixed. For example: if consumers are not flowing through after step 3 in the process, then there is something wrong for example is the next button not working or is the next step that you intend the consumers to take not clear enough. The funnel below shows it all…

In an ideal world, every advertiser likes to have a straight funnel i.e. all the consumers flowing straight through the checkout process as shown in the image below.

Net net the most important thing is to track as much as you can, to find out where the consumers are bailing in the checkout process and try to fix that. It might not be an easy fix and it can take a few tests to figure out the exact reason and then you can have your Advertising "$" working for you i.e. generating "$$$".

This is a very common problem seen with e-tailers who have a pretty complex checkout process which could involve more than a simple 5 step process. I have personally seen processes which are over 40+ steps and if there is no web analytics is place, then the advertising team is blind to where consumers are dropping off.

The solution is available fairly easily - use tools like Google Analytics, Webtrends, Omniture - you can define the complete consumer flow, look at visually where people are clicking or missing out on clicking, and then act. It's a circle which goes on - track, test, implement, track again :)

Friday, June 6, 2008

Post-Impression Attribution for Display Media

Often times while measuring an online advertising (Display/Online Banners) campaign only post-click activities or revenue are attributed to the campaign. One more factor which gets missed out is the post-impression activities and revenue.

Just to clarify the differences in post-click and post-impression, post-click activities are those which are the actions taken by consumers after they clicked on the Ad. Often times, consumers just view the ad and do not click on the ad but later on go to the website and take the action – this has been termed as Post-Impression activities.

I believe that the post-impression attribution should be higher for campaigns which are more branding as opposed to direct response. Post-impression action happens when the ad leaves an impression in consumers mind and the consumer comes back later to the site and takes the action. Thus, shows increase in brand awareness of the product. When the consumer clicks on the ad, that indicates that the consumer engaged with the brand and then took action.

There are many ways of calculating the post-impression attribution i.e. what percentage of the post-impression activities should be attributed to the campaign. One of the ways is to create a test and control group. For the control group, it needs to be made sure that they are not exposed to the Ad. The next step should be to see how many post-impression actions are taken by the control group. The lift in the conversion rate (post-impression actions/impressions) in the test (group exposed to the Ad) and control group is the post-impressions attribution. Essentially, in this methodology the control group conversion rate is the rate at which the consumers would convert with no exposure to the test Ad.

Similar methodologies have also been used in the past to establish the effect of marketing in retail stores.

Tuesday, June 3, 2008

Google Analytics vs. Active Meter

I currently use both these site tracking tools on my blog. One of the biggest differences I see is that in Active Meter I am able to get real time data. Whenever there is a visit on the site, the active meter shows the visit – also it includes if the visitor was unique or returning.

How the visitor was referred to my site, the location of the visitor and also if they used a search engine to get to my site – what was the phrase they put in the search engine.

Getting all this information real-time is awesome; I guess this basic analysis helps a lot.

On the other hand with Google Analytics there is more than a 24 hour lag in getting the report. I looked up this morning and still June 2, 2008 report was not available. In this day and age, I think we should be able to get a lot more frequent reports. At least some basic information should be available real-time.

I also use Pagealizer which provides the scroll length, click time, page view length etc. and all this information is real-time as well.

Sunday, June 1, 2008

Aarav is 5 weeks old and has a GOAL..

Site served vs. Third Party Served Ads: What is the difference?

In the online advertising world, it is very important to know how are the Ads served – site served or third party served (DART, Atlas etc). What is the difference really and why is it so important?

The difference is that when the ads are site served, it is not possible to the ad metrics like – impressions, clicks, post-click and post-impression activities. The metrics need to be provided by the publisher or the site serving the ad. There are some workarounds to the problem like using a click-command to get click and post-click activities via the third party ad server but being able to get impressions and post-impressions activities is not possible. Also, if the media buy is a CPM (Cost per 1,000 impressions) it is hard to get media spend, as it would be calculated based on the impressions. So, one of the drawbacks is NOT being able to easily get metrics.

The advantage of site serving is that the page load times are usually faster. Since, one layer of communication – site with the Ad server is removed.

The advantage of using third party ad servers like DART and Atlas is that the daily reporting is seamless. The reports for the previous day can be pulled the next day using the easy to use interface.

Most of the advertisers are now moving to third party ad serving solutions, so that the reporting is seamless and also being able to get post-impressions activities is a big advantage.


"PageAlizer" is a great tool useful for providing site analytics. I am sure there are a lot more out there (Webtrends, Omiture etc) providing the same information.

This tool provides how much time each visitor takes to click and where do they click. It also provides a visual map of the pages, how much did the visitors scroll down and where did they click. It helps understand the length of the page upto which consumers are scrolling down to. This is very important as if the content that you want the consumers to see is below where they are scrolling up to then they are missing what you want them to see and take action. Also, if the "Call to Action" button is below that scroll then they are missing that as well.

The report also provides the time till click.

I installed it for my blog and it is giving a lot of information.....